def convert(self, stream, options, file_ext, log, accelerators):
from ebook_converter.ebooks.metadata.toc import TOC
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
from ebook_converter.utils.zipfile import ZipFile
self.options = options
self.log = log
pages, images = [], []
toc = TOC()
if file_ext == 'pmlz':
log.debug('De-compressing content to temporary directory...')
with TemporaryDirectory('_unpmlz') as tdir:
zf = ZipFile(stream)
zf.extractall(tdir)
pmls = glob.glob(os.path.join(tdir, '*.pml'))
for pml in pmls:
html_name = os.path.splitext(
os.path.basename(pml))[0] + '.html'
html_path = os.path.join(os.getcwd(), html_name)
pages.append(html_name)
log.debug('Processing PML item %s...', pml)
ttoc = self.process_pml(pml, html_path)
toc += ttoc
images = self.get_images(stream, tdir, True)
else:
toc = self.process_pml(stream, 'index.html')
pages.append('index.html')
if hasattr(stream, 'name'):
images = self.get_images(
stream, os.path.abspath(os.path.dirname(stream.name)))
# We want pages to be orded alphabetically.
pages.sort()
manifest_items = []
for item in pages + images:
manifest_items.append((item, None))
from ebook_converter.ebooks.metadata.meta import get_metadata
log.debug('Reading metadata from input file...')
mi = get_metadata(stream, 'pml')
if 'images/cover.png' in images:
mi.cover = 'images/cover.png'
opf = OPFCreator(os.getcwd(), mi)
log.debug('Generating manifest...')
opf.create_manifest(manifest_items)
opf.create_spine(pages)
opf.set_toc(toc)
with open('metadata.opf', 'wb') as opffile:
with open('toc.ncx', 'wb') as tocfile:
opf.render(opffile, tocfile, 'toc.ncx')
return os.path.join(os.getcwd(), 'metadata.opf')
def create_opf(self, output_dir, images, toc):
with directory.CurrentDir(output_dir):
if 'cover.png' in images:
self.mi.cover = os.path.join('images', 'cover.png')
opf = OPFCreator(output_dir, self.mi)
manifest = [('index.html', None)]
for i in images:
manifest.append((os.path.join('images', i), None))
opf.create_manifest(manifest)
opf.create_spine(['index.html'])
opf.set_toc(toc)
with open('metadata.opf', 'wb') as opffile:
with open('toc.ncx', 'wb') as tocfile:
opf.render(opffile, tocfile, 'toc.ncx')
return os.path.join(output_dir, 'metadata.opf')
def create_opf(self, output_dir, images):
with directory.CurrentDir(output_dir):
opf = OPFCreator(output_dir, self.mi)
manifest = [('index.html', None)]
for i in images:
manifest.append((os.path.join('images/', i), None))
opf.create_manifest(manifest)
opf.create_spine(['index.html'])
with open('metadata.opf', 'wb') as opffile:
opf.render(opffile)
return os.path.join(output_dir, 'metadata.opf')
def __call__(self, stream, odir, log):
from ebook_converter.utils.zipfile import ZipFile
from ebook_converter.ebooks.metadata.odt import get_metadata
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
if not os.path.exists(odir):
os.makedirs(odir)
with directory.CurrentDir(odir):
log.info('Extracting ODT file...')
stream.seek(0)
mi = get_metadata(stream, 'odt')
if not mi.title:
mi.title = 'Unknown'
if not mi.authors:
mi.authors = ['Unknown']
self.filter_load(stream, mi, log)
# NOTE(gryf): Here is a workaround for ODF2XHTML.xhtml() method,
# which expects, that all lines are strings.
html = ''.join([str(l) for l in self.lines])
# A blanket img specification like this causes problems
# with EPUB output as the containing element often has
# an absolute height and width set that is larger than
# the available screen real estate
html = html.replace('img { width: 100%; height: 100%; }', '')
# odf2xhtml creates empty title tag
html = html.replace('<title></title>',
'<title>%s</title>' % (mi.title, ))
try:
html = self.fix_markup(html, log)
except:
log.exception('Failed to filter CSS, conversion may be slow')
with open('index.xhtml', 'wb') as f:
f.write(polyglot.as_bytes(html))
zf = ZipFile(stream, 'r')
self.extract_pictures(zf)
opf = OPFCreator(os.path.abspath(os.getcwd()), mi)
opf.create_manifest([(os.path.abspath(os.path.join(r, f2)), None)
for r, _, fnames in os.walk(os.getcwd())
for f2 in fnames])
opf.create_spine([os.path.abspath('index.xhtml')])
with open('metadata.opf', 'wb') as f:
opf.render(f)
return os.path.abspath('metadata.opf')
def write(self, doc):
toc = create_toc(doc, self.body, self.resolved_link_map, self.styles,
self.object_map, self.log, self.namespace)
raw = html.tostring(self.html,
encoding='utf-8',
doctype='<!DOCTYPE html>')
with open(os.path.join(self.dest_dir, 'index.html'), 'wb') as f:
f.write(raw)
css = self.styles.generate_css(self.dest_dir, self.docx,
self.notes_nopb, self.nosupsub)
if css:
with open(os.path.join(self.dest_dir, 'docx.css'), 'wb') as f:
f.write(css.encode('utf-8'))
opf = OPFCreator(self.dest_dir, self.mi)
opf.toc = toc
opf.create_manifest_from_files_in([self.dest_dir])
for item in opf.manifest:
if item.media_type == 'text/html':
item.media_type = mimetypes.guess_type('a.xhtml')[0]
opf.create_spine(['index.html'])
if self.cover_image is not None:
opf.guide.set_cover(self.cover_image)
def process_guide(E, guide):
if self.toc_anchor is not None:
guide.append(
E.reference(href='index.html#' + self.toc_anchor,
title='Table of Contents',
type='toc'))
toc_file = os.path.join(self.dest_dir, 'toc.ncx')
with open(os.path.join(self.dest_dir, 'metadata.opf'),
'wb') as of, open(toc_file, 'wb') as ncx:
opf.render(of, ncx, 'toc.ncx', process_guide=process_guide)
if os.path.getsize(toc_file) == 0:
os.remove(toc_file)
return os.path.join(self.dest_dir, 'metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators):
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
from ebook_converter.ebooks.pdf.pdftohtml import pdftohtml
log.debug('Converting file to html...')
# The main html file will be named index.html
self.opts, self.log = options, log
if options.new_pdf_engine:
return self.convert_new(stream, accelerators)
pdftohtml(os.getcwd(), stream.name, options.no_images)
from ebook_converter.ebooks.metadata.meta import get_metadata
log.debug('Retrieving document metadata...')
mi = get_metadata(stream, 'pdf')
opf = OPFCreator(os.getcwd(), mi)
manifest = [('index.html', None)]
images = os.listdir(os.getcwd())
images.remove('index.html')
for i in images:
manifest.append((i, None))
log.debug('Generating manifest...')
opf.create_manifest(manifest)
opf.create_spine(['index.html'])
log.debug('Rendering manifest...')
with open('metadata.opf', 'wb') as opffile:
opf.render(opffile)
if os.path.exists('toc.ncx'):
ncxid = opf.manifest.id_for_path('toc.ncx')
if ncxid:
with open('metadata.opf', 'r+b') as f:
raw = f.read().replace(
b'<spine',
b'<spine toc="%s"' % polyglot.as_bytes(ncxid))
f.seek(0)
f.write(raw)
return os.path.join(os.getcwd(), 'metadata.opf')
def opf_writer(path, opf_name, manifest, spine, mi):
opf = OPFCreator(path, mi)
opf.create_manifest(manifest)
opf.create_spine(spine)
with open(os.path.join(path, opf_name), 'wb') as opffile:
opf.render(opffile)
def convert(self, stream, opts, file_ext, log, accelerators):
from ebook_converter.ebooks.metadata import MetaInformation
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
from ebook_converter.ebooks.metadata.toc import TOC
self.opts, self.log = opts, log
if file_ext == 'cbc':
comics_ = self.get_comics_from_collection(stream)
else:
comics_ = [['Comic', os.path.abspath(stream.name)]]
stream.close()
comics = []
for i, x in enumerate(comics_):
title, fname = x
cdir = 'comic_%d' % (i + 1) if len(comics_) > 1 else '.'
cdir = os.path.abspath(cdir)
if not os.path.exists(cdir):
os.makedirs(cdir)
pages = self.get_pages(fname, cdir)
if not pages:
continue
if self.for_viewer:
comics.append(
(title, pages, [self.create_viewer_wrapper(pages)]))
else:
wrappers = self.create_wrappers(pages)
comics.append((title, pages, wrappers))
if not comics:
raise ValueError('No comic pages found in %s' % stream.name)
mi = MetaInformation(
os.path.basename(stream.name).rpartition('.')[0], ['Unknown'])
opf = OPFCreator(os.getcwd(), mi)
entries = []
def href(x):
if len(comics) == 1:
return os.path.basename(x)
return '/'.join(x.split(os.sep)[-2:])
cover_href = None
for comic in comics:
pages, wrappers = comic[1:]
page_entries = [(x, None) for x in map(href, pages)]
entries += [(w, None) for w in map(href, wrappers)] + page_entries
if cover_href is None and page_entries:
cover_href = page_entries[0][0]
opf.create_manifest(entries)
spine = []
for comic in comics:
spine.extend(map(href, comic[2]))
self._images = []
for comic in comics:
self._images.extend(comic[1])
opf.create_spine(spine)
if self.for_viewer and cover_href:
opf.guide.set_cover(cover_href)
toc = TOC()
if len(comics) == 1:
wrappers = comics[0][2]
for i, x in enumerate(wrappers):
toc.add_item(href(x), None, 'Page %d' % (i + 1), play_order=i)
else:
po = 0
for comic in comics:
po += 1
wrappers = comic[2]
stoc = toc.add_item(href(wrappers[0]),
None,
comic[0],
play_order=po)
if not opts.dont_add_comic_pages_to_toc:
for i, x in enumerate(wrappers):
stoc.add_item(href(x),
None,
'Page %d' % (i + 1),
play_order=po)
po += 1
opf.set_toc(toc)
with open('metadata.opf', 'wb') as m, open('toc.ncx', 'wb') as n:
opf.render(m, n, 'toc.ncx')
return os.path.abspath('metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators):
from ebook_converter.ebooks.metadata.meta import get_metadata
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
from ebook_converter.ebooks.rtf2xml.ParseRtf import \
RtfInvalidCodeException
from ebook_converter.ebooks.rtf.input import InlineClass
self.opts = options
self.log = log
self.log('Converting RTF to XML...')
try:
xml = self.generate_xml(stream.name)
except RtfInvalidCodeException as e:
self.log.exception('Unable to parse RTF')
raise ValueError('This RTF file has a feature calibre does not '
'support. Convert it to HTML first and then try '
'it.\n%s' % e)
d = glob.glob(os.path.join('*_rtf_pict_dir', 'picts.rtf'))
if d:
imap = {}
try:
imap = self.extract_images(d[0])
except Exception:
self.log.exception('Failed to extract images...')
self.log('Parsing XML...')
doc = etree.fromstring(xml)
border_styles = self.convert_borders(doc)
for pict in doc.xpath(
'//rtf:pict[@num]',
namespaces={'rtf': 'http://rtf2xml.sourceforge.net/'}):
num = int(pict.get('num'))
name = imap.get(num, None)
if name is not None:
pict.set('num', name)
self.log('Converting XML to HTML...')
inline_class = InlineClass(self.log)
with open(
pkg_resources.resource_filename('ebook_converter',
'data/rtf.xsl')) as fobj:
styledoc = etree.fromstring(fobj.read())
extensions = {('calibre', 'inline-class'): inline_class}
transform = etree.XSLT(styledoc, extensions=extensions)
result = transform(doc)
html = u'index.xhtml'
with open(html, 'wb') as f:
res = as_bytes(transform.tostring(result))
# res = res[:100].replace('xmlns:html', 'xmlns') + res[100:]
# clean multiple \n
res = re.sub(b'\n+', b'\n', res)
# Replace newlines inserted by the 'empty_paragraphs' option in
# rtf2xml with html blank lines
# res = re.sub('\s*<body>', '<body>', res)
# res = re.sub('(?<=\n)\n{2}',
# u'<p>\u00a0</p>\n'.encode('utf-8'), res)
f.write(res)
self.write_inline_css(inline_class, border_styles)
stream.seek(0)
mi = get_metadata(stream, 'rtf')
if not mi.title:
mi.title = 'Unknown'
if not mi.authors:
mi.authors = ['Unknown']
opf = OPFCreator(os.getcwd(), mi)
opf.create_manifest([(u'index.xhtml', None)])
opf.create_spine([u'index.xhtml'])
opf.render(open(u'metadata.opf', 'wb'))
return os.path.abspath(u'metadata.opf')
def convert(self, stream, options, file_ext, log, accelerators):
from ebook_converter.ebooks.metadata.fb2 import ensure_namespace
from ebook_converter.ebooks.metadata.fb2 import get_fb2_data
from ebook_converter.ebooks.metadata.opf2 import OPFCreator
from ebook_converter.ebooks.metadata.meta import get_metadata
from ebook_converter.ebooks.chardet import xml_to_unicode
self.log = log
log.debug('Parsing XML...')
raw = get_fb2_data(stream)[0]
raw = raw.replace(b'\0', b'')
raw = xml_to_unicode(raw,
strip_encoding_pats=True,
assume_utf8=True,
resolve_entities=True)[0]
try:
doc = etree.fromstring(raw)
except etree.XMLSyntaxError:
doc = etree.fromstring(raw.replace('& ', '&'))
if doc is None:
raise ValueError('The FB2 file is not valid XML')
doc = ensure_namespace(doc)
try:
fb_ns = doc.nsmap[doc.prefix]
except Exception:
fb_ns = FB2NS
NAMESPACES = {'f': fb_ns, 'l': const.XLINK_NS}
stylesheets = doc.xpath('//*[local-name() = "stylesheet" and '
'@type="text/css"]')
css = ''
for s in stylesheets:
css += etree.tostring(
s, encoding='unicode', method='text', with_tail=False) + '\n\n'
if css:
import css_parser
import logging
parser = css_parser.CSSParser(fetcher=None,
log=logging.getLogger('calibre.css'))
XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % const.XHTML_NS
text = XHTML_CSS_NAMESPACE + css
log.debug('Parsing stylesheet...')
stylesheet = parser.parseString(text)
stylesheet.namespaces['h'] = const.XHTML_NS
css = stylesheet.cssText
if isinstance(css, bytes):
css = css.decode('utf-8', 'replace')
css = css.replace('h|style', 'h|span')
css = re.sub(r'name\s*=\s*', 'class=', css)
self.extract_embedded_content(doc)
log.debug('Converting XML to HTML...')
with open(
pkg_resources.resource_filename('ebook_converter',
'data/fb2.xsl')) as f:
ss = f.read()
ss = ss.replace("__FB_NS__", fb_ns)
if options.no_inline_fb2_toc:
log.info('Disabling generation of inline FB2 TOC')
ss = re.compile(r'<!-- BUILD TOC -->.*<!-- END BUILD TOC -->',
re.DOTALL).sub('', ss)
styledoc = etree.fromstring(ss)
transform = etree.XSLT(styledoc)
result = transform(doc)
# Handle links of type note and cite
notes = {
a.get('href')[1:]: a
for a in result.xpath('//a[@link_note and @href]')
if a.get('href').startswith('#')
}
cites = {
a.get('link_cite'): a
for a in result.xpath('//a[@link_cite]') if not a.get('href', '')
}
all_ids = {x for x in result.xpath('//*/@id')}
for cite, a in cites.items():
note = notes.get(cite, None)
if note:
c = 1
while 'cite%d' % c in all_ids:
c += 1
if not note.get('id', None):
note.set('id', 'cite%d' % c)
all_ids.add(note.get('id'))
a.set('href', '#%s' % note.get('id'))
for x in result.xpath('//*[@link_note or @link_cite]'):
x.attrib.pop('link_note', None)
x.attrib.pop('link_cite', None)
for img in result.xpath('//img[@src]'):
src = img.get('src')
img.set('src', self.binary_map.get(src, src))
index = transform.tostring(result)
with open('index.xhtml', 'wb') as f:
f.write(index.encode('utf-8'))
with open('inline-styles.css', 'wb') as f:
f.write(css.encode('utf-8'))
stream.seek(0)
mi = get_metadata(stream, 'fb2')
if not mi.title:
mi.title = 'Unknown'
if not mi.authors:
mi.authors = ['Unknown']
cpath = None
if mi.cover_data and mi.cover_data[1]:
with open('fb2_cover_calibre_mi.jpg', 'wb') as f:
f.write(mi.cover_data[1])
cpath = os.path.abspath('fb2_cover_calibre_mi.jpg')
else:
for img in doc.xpath('//f:coverpage/f:image',
namespaces=NAMESPACES):
href = img.get('{%s}href' % const.XLINK_NS,
img.get('href', None))
if href is not None:
if href.startswith('#'):
href = href[1:]
cpath = os.path.abspath(href)
break
opf = OPFCreator(os.getcwd(), mi)
entries = [(f2, mimetypes.guess_type(f2)[0])
for f2 in os.listdir(u'.')]
opf.create_manifest(entries)
opf.create_spine(['index.xhtml'])
if cpath:
opf.guide.set_cover(cpath)
with open('metadata.opf', 'wb') as f:
opf.render(f)
return os.path.join(os.getcwd(), 'metadata.opf')
def create_opf(self, htmlfile, guide=None, root=None):
mi = getattr(self.book_header.exth, 'mi', self.embedded_mi)
if mi is None:
mi = MetaInformation(self.book_header.title, ['Unknown'])
opf = OPFCreator(os.path.dirname(htmlfile), mi)
if hasattr(self.book_header.exth, 'cover_offset'):
opf.cover = 'images/%05d.jpg' % (self.book_header
.exth.cover_offset + 1)
elif mi.cover is not None:
opf.cover = mi.cover
else:
opf.cover = 'images/%05d.jpg' % 1
if not os.path.exists(os.path.join(os.path.dirname(htmlfile),
* opf.cover.split('/'))):
opf.cover = None
cover = opf.cover
cover_copied = None
if cover is not None:
cover = cover.replace('/', os.sep)
if os.path.exists(cover):
ncover = 'images'+os.sep+'calibre_cover.jpg'
if os.path.exists(ncover):
os.remove(ncover)
shutil.copyfile(cover, ncover)
cover_copied = os.path.abspath(ncover)
opf.cover = ncover.replace(os.sep, '/')
manifest = [(htmlfile, 'application/xhtml+xml'),
(os.path.abspath('styles.css'), 'text/css')]
bp = os.path.dirname(htmlfile)
added = set()
for i in getattr(self, 'image_names', []):
path = os.path.join(bp, 'images', i)
added.add(path)
manifest.append((path,
mimetypes.guess_type(path)[0] or 'image/jpeg'))
if cover_copied is not None:
manifest.append((cover_copied, 'image/jpeg'))
opf.create_manifest(manifest)
opf.create_spine([os.path.basename(htmlfile)])
toc = None
if guide is not None:
opf.create_guide(guide)
for ref in opf.guide:
if ref.type.lower() == 'toc':
toc = ref.href()
ncx_manifest_entry = None
if toc:
ncx_manifest_entry = 'toc.ncx'
elems = root.xpath('//*[@id="%s"]' % toc.partition('#')[-1])
tocobj = None
ent_pat = re.compile(r'&(\S+?);')
if elems:
tocobj = TOC()
found = False
reached = False
for x in root.iter():
if x == elems[-1]:
reached = True
continue
if reached and x.tag == 'a':
href = x.get('href', '')
if href and re.match(r'\w+://', href) is None:
try:
text = ' '.join([t.strip() for t in
x.xpath('descendant:'
':text()')])
except Exception:
text = ''
text = ent_pat.sub(entities.entity_to_unicode,
text)
item = tocobj.add_item(toc.partition('#')[0],
href[1:], text)
item.left_space = int(self.get_left_whitespace(x))
found = True
if (reached and found and
x.get('class', None) == 'mbp_pagebreak'):
break
if tocobj is not None:
tocobj = self.structure_toc(tocobj)
opf.set_toc(tocobj)
return opf, ncx_manifest_entry
def write_opf(self, guide, toc, spine, resource_map):
mi = self.header.exth.mi
if (self.cover_offset is not None
and self.cover_offset < len(resource_map)):
mi.cover = resource_map[self.cover_offset]
if len(list(toc)) < 2:
self.log.warning('KF8 has no metadata Table of Contents')
for ref in guide:
if ref.type == 'toc':
href = ref.href()
href, frag = urllib.parse.urldefrag(href)
if os.path.exists(href.replace('/', os.sep)):
try:
toc = self.read_inline_toc(href, frag)
except Exception:
self.log.exception('Failed to read inline ToC')
opf = OPFCreator(os.getcwd(), mi)
opf.guide = guide
def exclude(path):
return os.path.basename(path) == 'debug-raw.html'
# If there are no images then the azw3 input plugin dumps all
# binary records as .unknown images, remove them
if (self.for_tweak and os.path.exists('images')
and os.path.isdir('images')):
files = os.listdir('images')
unknown = [x for x in files if x.endswith('.unknown')]
if len(files) == len(unknown):
[os.remove('images/' + f) for f in files]
if self.for_tweak:
try:
os.remove('debug-raw.html')
except Exception:
pass
opf.create_manifest_from_files_in([os.getcwd()], exclude=exclude)
for entry in opf.manifest:
if entry.mime_type == 'text/html':
entry.mime_type = 'application/xhtml+xml'
opf.create_spine(spine)
opf.set_toc(toc)
ppd = getattr(self.header.exth, 'page_progression_direction', None)
if ppd in {'ltr', 'rtl', 'default'}:
opf.page_progression_direction = ppd
pwm = getattr(self.header.exth, 'primary_writing_mode', None)
if pwm is not None:
opf.primary_writing_mode = pwm
with open('metadata.opf', 'wb') as of, open('toc.ncx', 'wb') as ncx:
opf.render(of, ncx, 'toc.ncx')
return 'metadata.opf'